Zurich Archives Push to Purge Duplicate Images From City's Digital Records — Here's Where Things Stand This Week

Stadtarchiv Zürich, the city's official records and image repository on Neumarkt, confirmed this week that a joint deduplication sweep covering roughly 340,000 digital image files is now in its final verification phase. The project — run in partnership with Zentralbibliothek Zürich on Zähringerplatz and the Bildarchiv at the Museum für Gestaltung on Ausstellungsstrasse — has been running since March 2026 and is due to close out before the end of July.

The timing matters. Swiss federal archiving standards updated in early 2025 now require cantonal institutions receiving public funding to demonstrate measurable data efficiency targets by the end of the 2026 fiscal year. Duplicate image files — the same photograph stored in multiple formats, resolutions, or locations across networked servers — inflate storage costs, slow retrieval systems, and make metadata tagging inconsistent. For institutions that also serve researchers at ETH Zürich and the Universität Zürich on Rämistrasse, messy archives translate directly into slower academic workflows.

What the Week's Audit Turned Up

The sweep identified a duplication rate of approximately 18 percent across the combined holdings — meaning roughly one in five stored image files was a functional copy of another already in the system. That figure, confirmed by Stadtarchiv, is higher than the consortium initially projected when it budgeted the project at CHF 280,000 in the February 2026 cantonal supplementary budget submission. The overrun risk, according to the project documentation reviewed by The Daily Zurich, stems largely from format fragmentation: files originally digitised at different resolutions by different departments were not flagged as duplicates by the standard hash-matching software, and required a secondary manual review layer.

The Museum für Gestaltung's Bildarchiv alone contributed around 47,000 images to the shared pool. A significant portion dated from analogue-to-digital conversion work carried out between 2009 and 2015, when Swiss cultural institutions received federal co-funding under a digitisation stimulus programme. Those older files carried inconsistent EXIF metadata, complicating automated sorting. Staff at Zentralbibliothek spent much of the past week manually cross-checking roughly 6,200 flagged near-duplicates — images where pixel content matched but file names, dates, or attribution fields differed.

The practical stakes extend beyond server tidiness. Zurich's housing shortage and rapid densification have made neighbourhood planning documents and historical street-level photography increasingly valuable for appeals and rezoning disputes under cantonal law. The Wohnungsnot crisis has pushed planning offices to cross-reference decades of urban imagery. Duplicate or mislabelled files can introduce errors into those processes, and in at least two recent cases before the Baurekurskommission, archival image discrepancies were cited in procedural objections — though the commission did not rule on evidentiary grounds in either instance.

What Comes Next for Researchers and the Public

Once the verification phase closes, the consortium plans to publish a cleaned, unified image index accessible through e-rara.ch, the Swiss digital platform used by libraries and archives across the German-speaking cantons. The target date is 31 July 2026. Files confirmed as genuine duplicates will not be deleted outright; instead, they will be moved to a cold-storage tier, retaining preservation copies while removing them from active search indexing.

For members of the public using the Stadtarchiv reading room on Neumarkt, the practical change will be visible in the online catalogue: search results for historical Zurich street photography are currently returning multiple near-identical results for the same negative, particularly for Langstrasse and Niederdorf imagery from the 1960s and 1970s. After the index is updated, those results should consolidate into single canonical entries with linked provenance notes.

Researchers affiliated with ETH Zürich's Institute for the History and Theory of Architecture, which draws heavily on the Bildarchiv holdings, have been briefed on the transition timeline. The deduplication work has also prompted a broader conversation among the three institutions about adopting a shared cataloguing standard going forward, which would prevent the problem from recurring as new digitisation projects come online in 2027.